The goal of the proposed research is to improve the noninvasive identification of patients with ischemi heart disease vulnerable to ventricular tachycardia (VT) or ventricular fibrillation (VF). Analyses of 3-dimensional intraoperative and body surface maps of ventricular activation in patients indicate that current methods of assessing risk fail to detect completely the electrophysiologic abnormalities induced during ventricular infarct healing that increase susceptibility to VT/VF. Recent findings demonstrate that: VT initiates at infarct border zones by intramural reentry and focal mechanisms; tissue generating nonsustained VT is not a surrogate for sustained VT; analysis limited to the terminal QRS complex of signal-averaged ECGs excludes detection of more than 95% of the signals generated by myocardium critical for VT; and during the entire cardiac cycle, there are previously unrecognized spectral, temporal, and spatial features of body surface and inferred epicardial potentials that distinguish patients with VT. Based on these findings, the next step toward improving risk stratification is a focus on bioelectrical signals generated by myocardium enveloping arrhythmogenic tissue. The investigators will characterize the spectral and temporal features that distinguish epicardial potentials inferred over arrhythmogenic loci in infarct zones. They propose then to establish that these bioelectric signals are a more complete fingerprint of the electrophysiologic derangements that increase susceptibility to VT/VF by testing the hypothesis that detection of the uncovered abnormal signals improves identification of patients vulnerable to sustained ventricular arrhythmias. Epicardial potentials subtending arrhythmogenic and non-arrhythmogenic infarct loci (defined ultrasonically and by catheter endocardial mapping) and normal myocardium will be inferred from 190-lead body surface maps, individualized heart-torso models, and echocardiographic images from patients with (n=40) and without (n=40) VT or VF. Magnitude, phase and group-delay spectra of potentials will be analyzed to determine the frequency band(s) and their temporal location(s) in the cardiac cycle that distinguish signals generated locally by myocardium enveloping arrhythmogenic tissue. Fina studies performed in 100 subjects will test whether detection of this fingerprint, developed based on an increased understanding of the pathophysiologic basis of VT in humans, improves identification of patients with prior myocardial infarction inducible into sustained VT.